Variation in Breeding Systems and Consequences for Reproductive Traits in Erica

Abstract

Erica makes up 7% of all species in the Cape flora. It is the most species-rich genus in

the Cape Floristic Region (CFR), which has an area 0f 90 000km2. Erica species have

great inter- and intra-specific variation in floral form as well as in post-fire regeneration

strategies. Previous studies of other plant groups (Barrett et al. 1996; Button et al.

2012) have illustrated changes in floral traits with a shift from outcrossing to selfing.

The aim of this thesis was to determine whether similar changes occur in Erica. I thus

analysed breeding systems in Erica in relation to floral traits, pollinators, and fire

survivals strategies.

Seed and fruit set data obtained from hand-pollination treatments suggested that selffertilization

occurs rarely. Autonomous selfing did not generally yield significant seed

set. When comparing changes in floral traits with increased selfing ability and

pollination syndromes weak relationships were observed. Average plant height showed

a positive relationship with selfing ability but this correlation was not significant.

Comparing traits within small and large flowered species separately yielded significant

relationships between corolla size and selfing ability of small-flowered species..

Furthermore, general trends indicate that small-flowered species have increased ability

to self-fertilise when corolla sizes are larger, corolla apertures larger and herkogamy

reduced. Large-flowered species employ the same strategy but average corolla aperture

is reduced. A trade-off between the size and the number of flowers was seen within the

Erica genus. Small-flowered species had significantly more flowers compared to largerflowered

species that had markedly fewer flowers per unit height. This finding has

implications for the selfing potential of small-flowered species as increased selfincompatibility

may have evolved in order to reduce the effects of increased

geitonogamy due to increased floral number.

The prediction that self-fertilisation would be increased in seeders compared to

resprouters, on the basis that seeders are more reliant on seeds for persistence than

resprouters, was not realised when comparing the selfing ability of different firesurvival

strategies. Erica mammosa, a species with morphs possessing both fire-survival

strategies, shows no significant differences in selfing ability, this includes differences in

pollen-ovule (P/O) ratios. However, indices suggest the resprouting form of E. mammosa to have an increased ability to self-fertilize while the seeder form has an

increased ability for autonomous selfing. The 29 species analysed were divided into

outcrossers and facultative outcrossers based on selfing indices but these did not fit

neatly within Cruden’s proposed P/O ranges (facultative autogamy: P/Os= 32-397;

facultative outcrossers: P/Os= 160.7 - 2258.6; outcrossers: P/Os= 1062 - 19525). This

may be due to his classification of species into breeding systems being based on

relatively few distantly related species per category with extremely variable P/O ratios

per category. For example, outcrossers ranged an order of magnitude (from 1000+ to

20 000). It is also true that this ratio can be influenced by a variety of different factors,

these include: habitat, evolutionary history and pollination syndrome. Consequently,

P/O ratios in Erica do not seem to reflect pollination syndromes very well. Presumably,

sex allocation theory may explain the relationship of breeding system with P/O ratios

better.

Histological studies of pollen tube growth for self- and cross-pollinated flowers of eight

species suggested that Erica has late acting self-incompatibility (LSI). LSI is a barrier to

selfing that occurs in the ovary. However, I could not determine if the rejection process

occurs pre- or post-fertilization.

Although, a large amount of knowledge is still lacking, this preliminary study provided

insight into the reproductive biology of Erica.